Studies show that elderly persons residing in rural areas receive significantly less outpatient physician care than do urban elderly residents. Although the differentials appear to be large and persistent, surprisingly little research has been directed towards understanding their root causes or assessing their impact on outpatient treatments that require physician intervention including specialty referrals and the use of prescription drugs. This study proposes a formal assessment of the determinants of urban/rural differentials in ambulatory physician contacts and medicine use during 1992 for a nationally representative sample of community-dwelling elderly Medicare beneficiaries. This project will use a database that links person-level information from the new Medicare Current Beneficiary Survey (MCBS) with Medicare claims data and place-of-residence characteristics from the Area Resource File (ARF) and Census Bureau files. This project aims to measure the urban/rural utilization differentials in both probability of use and level of use among elderly persons for outpatient physician contacts (by site, physician specialty, referral status, and medical reason for the visit) and prescription drug use (by medical reason for the Rx and number of scripts per physician contact) during 1992. In addition, the project plans to determine the sensitivity of the differentials to alternative urban/rural typologies. The project will employ a hierarchical regression procedure based upon Andersen's behavioral model of health services utilization to explain the source of those differentials, making use of extensive data on the place of residence and the distance-to-care in the linked files. Finally, a series of regression decompositions and simulation exercises will be used to explore the impact of policy changes on health services utilization by rural elderly persons.